Intelligence in a modern insurgency: the case of the Maoist insurgency in Nepal

ABSTRACT Outside some well-known movements like Al Qaeda, there is little understanding of how insurgent movements in the Global South gather, process and manage intelligence. This paper is based on fieldwork in Nepal with former members of the Maoist Army. The Maoists fought a secretive insurgency war for ten years, signing a peace agreement in 2006. Fieldwork involved former combatants, intelligence officers and Maoist cadres and analyses the intelligence methodology of the Maoist insurgency, placing this in to the context of Nepal Government operations. The Maoists benefited from poor opponents but they did establish an effective system of intelligence into operations.


Introduction
The literature on intelligence rarely addresses intelligence structures and use within insurgent groups. 1 This is for a series of good reasons, not least because insurgent groups tend to be very secretive generally, and intelligence tends to be particularly sensitive, but also they also tend to lack formal, codified and documented systems that can be analysed by researchers after the events. Intelligence operators can also be extremely difficult to reach and are rarely willing to talk on record about previous experience of intelligence gathering and use, particularly where post-conflict measures like transitional justice are in operation. As a result, the intelligence literature rarely analyses comparative intelligence services within insurgent groups or the demand or usage of intelligence by those groups. 2,3 Geographically, this has also been reinforced by a bias in intelligence writing towards the Anglosphere, with an increasing number focussing on Soviet and post-Soviet space 4 and also the treatment of foreign intelligence services and non-state actors. However, Africa and South Asia are more or less absent from intelligence writing despite the fact that these geographical regions have been the focus of the vast majority of post-Second-World War violence.
This article therefore makes a significant contribution to the literature by describing the intelligence activities of the United Communist Party of Nepal (Maoist) (the 'Maoists') during their period of insurgency from 1996 to 2006. The article describes the intelligence activities of the Maoists and analyses how they were able to develop a critical advantage in many areas over their opponents and makes contribution to the existing literature on the complexity of insurgent intelligence operations in the global South, something that the extant literature is very reluctant to acknowledge. 5 Given the paucity of academic literature in this field, the paper follows the example of Gentry and Spencer in not concentrating on theorizing Maoist intelligence or rigorously comparing the Maoists with the FARC, but concentrating on the basic facts and using analysis of intelligence rooted in the theoretical literature: organisational structure; functions; consumers; external relationships; and performance. 6 The paper begins with a discussion about the methodology used in gathering data, which is always partial. However, the author was given extensive access to a wide range of intelligence actors from the period on both sides and over a period from 2009 until 2017. Consequently one element of the significance of this work is that it is based on first-hand accounts of intelligence actors themselves talking about the nuts and bolts of intelligence collection and their philosophy of intelligence. Despite the limitations of researching an area where triangulation is difficult, this adds credibility to the overall analysis.
Following a section on methodology, the paper then moves on to discuss the context of the insurgency itself, then concentrating on the intelligence organisation and infrastructure. It then goes through the core features of an intelligence mechanism placing the activities of the Maoist system in to a formal structure, governed by a Central Intelligence Bureau (CIB), and providing practical examples of activities where possible. Finally the paper comes to a conclusion that a series of contextual factorsnotably the failure of state intelligenceenhanced an already capable intelligence mechanism. The Maoist insurgents understood the importance of intelligence in their strategy, not just in terms of operational effectiveness, targeting and counter-intelligence, but also in developing an information strategy to make themselves more acceptable to the local population.

A note on methodology
The Maoists are notoriously secretive as an insurgent movement, even amongst the pantheon of insurgent movements. They have been very difficult to break down from the point of view of researchers, apart from at the village level, where some excellent work has been done on Maoist infiltration and relationships with local communities. 7 This work is based on research and policy work carried out over a number of years. The author first became involved in the Nepali peace process in 2008 as an external adviser to the Special Committee of the Nepali Constituent Assembly on the reintegration of the Maoist Combatants. This role continued until 2015. As an academic invited by former students to participate in the process, I was able to develop considerable trust with several Maoist leaders and members of the military and political groupings within the movement. 8 My role was as a facilitator and as someone who could act as an interpreter of the international community as well as someone who had worked in this area before and could bring international experience.
This greatly privileged position allowed me to develop a research project supported by the Government of Sweden that followed, interviewed and documented the transition to civilian life of around 250 former combatants across Nepal. This research project also provided me with access to a number of additional individuals who had been part of the Maoist intelligence network and I was then able to interview them about their involvement. In this way I managed to carry out around 20 interviews specifically on intelligence. This was also supplemented by a field trip to Dang and Rolpa Districts, accompanied by a former senior intelligence commander and involving four district intelligence chiefs, including along the border with India. 9 Additional interviews were carried out with senior military commanders, including former and current Chiefs of Army Staff and senior commanders (all of whom deployed as commanders during the war), as well as most of the Maoist High Command, and several middle level officials engaged in the day to day management of intelligence in the districts. This included all of the central intelligence group close to the command structure, and included, for example, Prachanda 10 (Chairman), Baldev (intelligence commander and member of the Standing Committee of the UCP(M)), Ananta (Divisional Commander and member of the Standing Committe), Nirmal GC (intelligence officer and Battalion Vice Commander) and Milan (Zonal intelligence Commander, Terai, and Prachanda's ADC). We will return to Baldev and Milan below.
I was also able to interview several Nepal Army commanders and former officers involved in military intelligence, who largely confirmed the Maoist narrative, along with providing additional information on the lack of intelligence they had on deploying to the field. 11 Therefore, I am confident that the paper provides an accurate, if incomplete, picture of Maoist intelligence during the war.

Maoist people's war in Nepal
In 1996 the Communist Party of Nepal (Maoist) declared the 'people's war' in Nepal. The ultra-left communist party declared this 'people's war' to establish a 'people's democracy' on 13 February 1996 by putting an end the multi-party democracy and constitutional monarchy that was established in 1990 following a popular movement. 12 The CPN (Maoist) started armed activities from the Hill districts of Nepal (Gorkha, Rolpa, Rukum, Salyan, Sindhuli, Jajarkot), mainly in the central belt in the West, by attacking local police posts and cadres of political opponents, in particular, the Nepali Congress (NC) and the Communist Party of Nepal (Unified Marxist Leninist) [CPN (UML)]. 13 In the aftermath of the initial armed attacks, the then government led by Nepali Congress President Girija Prasad Koirala decided to mobilise the Nepal Police to control the insurgency. The Nepal Police started by launching what have become two notorious operationsthe 'Romeo' and 'Kilo-Sera Two' operationsin those districts to control the Maoist's activities. Taking a very aggressive approach to counter-insurgency with an inexperienced and under-resourced force led to an excessive use of force, coupled with gross human rights violations by the Police involving violent suppression of local civilian populations on the pretext of controlling the Maoist's violent activities. This initial approach not only failed to establish control but also led to severe criticism of the security forces and managed to accelerate the rebellion through increasing support for the Maoists amongst much of the rural civilian population.
Notably the army itself did not take a leading role in the counter-insurgency effort until around four years before the end of the war. The civilian police who were deployed across the districts were backed up by various secret counter-insurgency units and special task forces, but most commonly by a new paramilitary force, the Armed Police Force (APF). Along with an occasionally heavy-handed approach to dealing with 'People's Enemies' by the Maoists this resulted in widespread suppression of the civilian population. The war itself produced several indirect casualties. A low intensity war, the Nepali conflict was largely perpetrated through the extensive use of IEDs and assassinations, but also widespread abductions and bombings, including of civilian vehicles. A notable feature of the war was also the use of violence and intimidation for extortion. Around 17,886 people lost their lives during the armed conflict and around 1,530 people disappeared in this decade, most of them during the state of emergency. 14 In addition, 79,571 persons were internally displaced during the armed conflict due to either Maoist violence or the counter operations of the security agencies. 15 A core tactic of the insurgency was to limit and reduce the influence and ability of the state at the local level to provide services and security to the local population. The destruction of VDC offices and other local state institutional buildings was accompanied by the extensive abduction of local government officers and sometimes their families. This degradation of the state outside the main urban centres contributed to rapid urbanisation in the Kathmandu Valley and the Terai Region creating significant problems for the Nepali Government in re-establishing control and service provision in the countryside. At the same time, the police, unable to defend isolated police posts, with no equipment and ill-prepared for counter-insurgency, retreated to defensible areas leaving large parts of the country without state security of any sort. 16 This resulted in a loss of confidence in the security services that severely hampered their ability to collect any intelligence outside urban areas and allowing the Maoists a virtual monopoly of intelligence in their core areas. In interviews with senior military commanders and politicians on the government side at this time, it is striking that they all say that they had virtually no intelligence at all on these areas, even when they were deploying.
The ten year insurgency was a relatively low intensity conflict that escalated in to larger scale violence. The beginning of the war was very much attacks by small groups of Maoist insurgents against isolated police posts, kidnappings, assassinations and the use of explosive devices against targets regarded as legitimate, particularly the civilian police, the APF and later, the army. Increasingly, however, the conflict escalated in to battles consisting of hundreds, if not more, insurgents attacking static defences of the security services. These larger attacks were extremely mixed, with more than 300 insurgents being killed as they attacked a military base at Khara in Rukum District, for example, but overrunning a poorly sited base at Pili in Karnali District. Eventually the Maoists tended to hold sway in the countryside and particularly in the hills, with urban areas of the Kathmandu Valley and the Terai being disputed and the Maoists being unable to effectively remove the security services from most district capitals or from bases like Bhalubang in Dang. 17 By the time of the battle of Bhalubang, the actual fighting had reached something of a stalemate. 18 The Maoist army was clearly unable to wage a full-sclae war because of lack of resources and significant outside support, whereas the Royal Nepal Army may have had some resources, but was poorly trained, lacked core counterinsurgency skillsnotably air power, sufficient special forces and intelligenceand also suffered from lack of popular support amongst the population. Government efforts were also characterised by poor relationships and coordination between services, at least partly as a result of interpersonal rivalries. This was not helped by the political crisis following the palace massacre and the replacement of a popular king with a very unpopular one and the inability of the political parties to work together. Eventually the political issues and royal takeovers in 2002 and 2005 built up in to popular pressure and an allparty alliance that allied with the Maoists to overthrow the regime. A ceasefire in 2005 was followed by a Comprehensive Peace Agreement in 2006 after 10 years of war and the Constituent Assembly was formed under a Republic, with the Maoists as the single biggest party.
The UCPN(M) is overtly Maoist in its approaches and tactical doctrine. At the time of its emergence from a well-established broad communist opposition in the early 1990s, it explicitly modelled itself on Sendero Luminoso (Shining Path), and took almost all political commentators by surprise. 19 'People's War was launched on 13 February 1996 and was viewed as an effective way to seize state power. Based on traditional communist rhetoric and approachesestablishment of a 'people's republic', an end to Indian imperialism, an end to capitalist exploitation and an end to an exploitative social hierarchy, including caste -Maoism was explicitly used as a methodological manual throughout the war. Since the system at the time was held to be 'reactionary' and governed by 'class enemies', people's war was held as the only way to bring this vision to fruition. Taking inspiration from Shining Path and also from the various Naxalite movements across parts of India, the insurgency was characterised by a mixture of infiltration in target areas and extreme violence in the service of 'elimination of class enemies'. 20 The Maoists 'progressive' and 'scientific' ideology was grafted on to a practical approach to People's War based around establishing links in a local community through infiltration and then developing a cadre-based clandestine governance infrastructure that may evolve in to the parallel structures within occupied areas with People's Courts and a functioning Maoist government. The Maoists claimed to be working towards a political solution, which for most of the war consisted of a negotiated settlement whereby non-Maoists would accede to their demands. An extensive clandestine governance network requires a military organisation to provide security. A Maoist military approach is well defined and falls in to three stages: defensive; stalemate and offensive. Essentially these stages are characterised by different approaches to the war itself. The defensive stage is characterised by terror and guerrilla tactics. The stalemate becomes more formal and involves mobile (or manoeuvre) warfare with roughly equivalent military formations battling with state forces, but not seeking to hold territory. Lastly, the 'war of position' sees a strategic offensive seeking to take and hold ground. In Nepal, the first phase was clearly defined, and certainly the military hierarchy felt they were entering the second phase with the launch of a general offensive in November 2001. However, for much of the war, the Maoists did effectively hold territory in their heartlands in the Mid-West even if they did not claim 'base areas'. In addition, whilst the war did enter a second phase, performance on formal battlefields tended to favour the Government forces, particularly when they could dig in. However, a stalemate was eventually reached at least partly because of the lack of initiative on the part of the government and their lack of agreed strategies and approaches. 21 At its core, however, the Maoist movement is a very tight knit group who were there from the very beginning. It may have a conventional military structure based on Chinese rankscompany commander, brigade commander and then generalbut the real cement that holds the movement together is the level of personal trust between the core of individuals, alongside a network of political commissars at all levels to ensure ideological commitment. The nature of the political and military command structure is frequently seen as a simple ideological control, although extensive discussions with officers of both types shows that there was certainly considerable pragmatism within these relationships and at some levels, where the military commander was also a senior member of the Party, there was felt to be no need for a commissar figure. 22 What is also clear is that whilst there was notionally a parallel structure of political and military officers, they frequently worked together and developed trust between each other and in some cases, were the same person. 23

Approaches to maoist intelligence
The Maoist Army established a Central Intelligence Bureau (CIB) that reported directly to the senior command of the army in the early days of the war. There were no written manuals or guidance for intelligence gathering. All intelligence networks needed some way of converting information in to intelligence and this was incorporated inside some individuals who were trusted and given individual authorities. CIB numbers are difficult to ascertain. Although a figure on 'around 100ʹ was mentioned in one interview, it was also pointed out that this number varied over time, but was also supplemented by people who had been trained and were linked to CIB, but who were operating as field officers. The CIB operated as a clearing house whose job was primarily to organise training and processing of information for the high command. Raw information came directly from the field, usually via district commanders or from intelligence operators. This information was then processed by CIB and then the finished intelligence passed on to the high command. Whilst this was by no means the only intelligence route, this system had been designed early in the insurgency and operated right through the conflict, and it is telling that the vast majority of the training carried out was in information processing rather than gathering, suggesting that the Maoists had recognised the importance of accurate intelligence. 24 CIB did establish a rudimentary training programme for intelligence cadres. Training was carried out within Nepal by those who were experienced in intelligence for a period of one month. I met several senior level personnel who had taken part in the training, but also more junior officers who had taken training as part of their role within the Maoist structure. There is no evidence of any foreign influence over the training per se and certainly not the direct training. The idea of this approach was to produce a small number of semi-professional intelligence cadres to operate within CIB as information processors, but also a group of operational intelligence commanders who could take local decisions but who could also identify what information could be used locally and what needed to be passed to the CIB.
Many of these were located at the centre of the organisation to provide intelligence support, but most were posted out to the divisional commands as field operators. This suggests an intelligence element to provide 'strategic intelligence' to leaders as well as smaller, lower-level intelligence groups that provided local political support, tactical military targeting, and perhaps counterintelligence. Many Maoist cadres, in keeping with Maoist doctrine, also contributed to intelligence, but the idea was that there would be an intelligence commander in each district to provide local intelligence, but also to act as a liaison with CIB and the centre. Decisions in general within the Maoists were highly centralised and doctrine was extremely important in guiding discipline of the movement. This also led to a form of organisation resembling the cell structures present in other insurgency movements and based on Mao's 'Organization of a Guerrilla Company' in his Appendix to On Guerrilla War. Whereby an intelligence officer forms one of the five executive members supporting the Company Commander, the Executive Officer and the Political Officer. This, in turn, is supplemented by the Maoist idea that everyone in the movement and indeed, in the area, is an intelligence asset.
The Maoist interpretation of that was to establish a centralised system of intelligence officers linked back to the CIB, but with clear guidelines about localised decisions and activities. A hard core of activists, sometimes from the very first revolutionary cell, formed an inner circle that also internalised intelligence as an activity. All six battalion commanders of the Maoist Army were drawn from this group, along with several Politburo members with special status, including Prachanga, Batterai, Ananta, Pasang, Prabhakar and Baldev amongst others, and they all had intelligence responsibilities at strategic level. All were trained in intelligence and some ran their own intelligence operations. Baldev ran an intelligence operation for two years from a photography shop in the Terai.
In practice, the Maoist system was a hybrid that combined both a separate, centralised intelligence operation and also involved some senior Maoists in intelligence gathering, which is unusual. At the same time, and following Maoist doctrine, these professional intelligence gathering activities were fed and supported by networks of local sympathisers in each area. In this way, Maoist intelligence became highly effective in rural areas where there was significant sympathy for the Maoists, but had issues and became increasingly reliant on professional intelligence gathering in the more urbanised areas of the Kathmandu Valley and the Terai region. In these areas, the Maoist intelligence operation was complicated by extensive vetting of potential volunteers and their families and enforcement of party discipline. 25 This structure also sought to overcome some of the central issues of all rural insurgency movements, namely the quality of recruits. Much like the FARC in Colombia, the Maoist military recruited mainly from rural areas amongst the most deprived sections of society. Recruiting large numbers of female combatants and Dalits meant that there was a shortage of educated recruits. To compensate for this, the Maoists initiated a series of educational policies within the movement, 26 but also fell back on to the structure built around the original members of the Maoist movement, many of whom were upper caste, including Batterai, Prachanda and Baldev (all Brahmin). It was this core that held the movement together and undertook a wide variety of activities as a narrow network of trusted leaders.
It should also be noted that the operational environment faced by the Maoists was largely positive since the Government intelligence system was effectively falling apart. The three main intelligence agencies were the Directorate of Military Intelligence (DMI), the Special Branch (police) and the National Investigation Department (NID) which reports to the Ministry of Home Affairs. However, with a few temporal exceptions, none of these talked to each other they rarely reported formally and there was little accountability. Consequently there were several instances where these agencies worked against each other and may even have provided information to the Maoists as part of this rivalry. Indeed the RNA officers interviewed, who had deployed during the war, were disparaging of the amount and quality of any intelligence they received. 27 Indeed, two senior officers both said that the first thing they had to do was to attempt to establish their own local intelligence networks because they had no information at all, whereas the Maoists knew everything about themhow many, which regiments, where they would be deployed and when. One officer described this as 'deploying blind'.

Intelligence organisation
Clearly the existence of a cadre of intelligence agents with some training within a CIB was designed to facilitate an improvement of the analysis of information and its transformation in to useful intelligence. However, the structure of the Maoist movement basically meant that the analysts and the consumers could be the same individuals but that overall levels of intelligence were significantly improved by a separate analytical capability. The practice of sending senior members to the field as the core of an intelligence network led to the situation whereby they would sift through information provided by district commanders and also from the networks themselves and then report to meetings where they themselves were present.
The role of district commanders also provides a separation of useful local information from the intelligence that is passed up the central chain to the high command. As a result, the system deployed was again both centralised and decentralised with local military information shared at the local level but a requirement for central approval meant that information sharing at least nominally took place. In this way, the Maoist structure resembles that of the Viet Cong and Viet Minh, but also Al Qaeda, Hezbollah and the Tamil Tigers rather than, say the decentralised structure of the FARC.
Basic field operations relied very much on the human structures of the Maoist organisation, particularly on the decision-making of district commanders. As such, the system was centralised and decentralised at the same time. Certain decisionsattacks, for examplewere centralised and taken by the Central Committee or in many instances under oversight of regional bureaus and command. However, District commanders were able to act with discretion over some particular aspects, e.g. certain types of assassination or robbery. It was up to the district chief to make decisions between important and less important information, which is why so many were senior, trusted cadres.
It is important to note that there were instances of abuse of this system with some much power vested in individuals. When district chiefs acted against the best interests of the Maoists, or were judged to have not passed on relevant decisions, or in fact when they were deemed to be following personal agendas, then they could be removed, suspended from their positions or demoted from the Central Committee to the District Committee or to the Area Committee, and so on. Of course, the acid test of this is whether any of them were actually removed, and in reality this appears to have been very common. In one interview, we discussed one specific District Chief (Akan) who was accused of acting against the Maoists and in his own interests. His particular crime was said to have been assassinating personal enemies, rather than those necessary to the cause, or those who had been sanctioned by the central Committee. Eventually he was removed and in his case, he was banned from the partythe most serious action short of execution. He was later assassinated by the military.

Intelligence collection and management
Mirroring the three stage structure of the People's War strategy, Krause divides intelligence management in to three distinct phases: organisational; guerrilla warfare; and conventional warfare. 28 However, as discussed above, this never happened in Nepal quite as cleanly as the model suggests. Having discussed the way in which Maoist intelligence was organised, there are questions about the practicality and management of intelligence itself and what the movement actually did with its information. The Maoists relied heavily on HUMINT 29 to support its intelligence activities. It had little or no imagery capability, although it was adept at stealing intelligence and other resources from the state forces, and Maoists did use mobile telecommunications and cameras to spread information and also to film barracks and installations prior to attack. 30 There were several examples of copied Government documents and training manuals being used by the Maoists as the basis for their own operations. Maoist officers, for example, followed a similar curriculum to the Nepal Army Officers because the materials were regularly stolen from Army HQ.
Internet access also provided additional sources of intelligence both through the wide use of telecommunications, the provision of links to the outside world, and also open source intelligence. Nevertheless it is also clear that the use of SIGINT was minor compared to HUMINT and the core of the CIB work consisted of management of informant networks, surveillance of potential targets and infiltration on different levels.

Informant networks
The most important element of intelligence gathering for the Maoists was undoubtedly through informant networks. To an extent, most Maoist fighters were trained to gather simple intelligence on the ground and feed it back through the centralised command structure, or through the intelligence cadres who frequently deployed in local areas. Every district across Nepal had an intelligence commander and varying numbers of subordinates. This is, in itself, an interesting feature of the strengths and weaknesses of this approach, not least because in strong Maoist areas, there was strong Maoist intelligence, whereas in the Terai and in the Kathmandu Valley it was much more dangerous to operate as intelligence gatherers or fighters.
This was exacerbated by the further reliance on local populations as a key source of intelligence, including wide networks of activists, sympathisers and also children. In fact, during an initial interview with a senior intelligence commander, he stated that the Maoist Army 'did not have intelligence, we just asked people', which is accurate within Maoist core areas and constitutes a good, ideological answer. 31 In strong Maoist areas and amongst particular ethnic groups like the Maghar population, there was excellent intelligence, but this diminished once the war moved in to the more mixed and mobile populations of the Terai, where more professional intelligence operations had to be established to compensate.

Surveillance
There was no formal counterintelligence unit, rather the Maoists preferred to rely on personal ties and structures, deploying systems of informants to undermine attempts made by the state to penetrate the Maoist networks. This was particularly effective in rural areas, where the state had effectively abdicated responsibility and where there were organised structures of resistance across many of the districts even before the war. Indeed, some of the attempts by the intelligence agencies to penetrate the Maoist heartlands were simply derisory.
One interviewee related a story about finding a former soldier who had been detailed to gather intelligence information on local Maoist operations. 32 In speaking to many Nepal Army commanders who were deployed to the countryside at that time, they uniformly state that they had no intelligence on local Maoist activities and they had to establish their own intelligence gathering operations. This unfortunate soldier was chosen from the deployed troops and sent out to the fields to pretend to farm and gather intelligence. However, he didn't know how to farmhe was from an urban backgroundand this was quickly spotted. The soldier was then executed.
The Maoists were not completely impervious to infiltration, however and their internal surveillance networks were breached before the battle of Bhalubang. The Armed Police Force (APF) garrison of the position had, in fact, been able to turn (and/or bribe) a Maoist company commander in to providing details of the upcoming attack on their position. 33 Although the commander was eventually executed once suspicions had been aroused. However, overall, in international terms the Maoists were remarkably closed and secretive and suffered no real high-level breaches in the ten years of war, which is quite remarkable. Even when one considers the poor record of the state security services in dealing with the Maoist threat, this is an impressive testimony to the high levels of internal trust and discipline within the movement.

Infiltration
Terrorists and insurgents who penetrate state institutions can derive at least five significant benefits 34 : information; opportunities to plant false information; stealing resources; identification of targets for bribery or recruitment; and, delegitimising the state's war effort. Infiltration is an old and highly effective target used by both insurgent groups and also state security services. There are contemporary examples of infiltration by Jamaat-e-Islami in Bangaldesh, Al Qaeda in the United Arab Emirates and of various insurgents in Iraq, as well as significant use of infiltration by the UK security services in Northern Ireland. 35 We do not really know the extent of Maoist penetration of the Nepali state, although we do know that at one stage a Government Home Minister was providing funds to the Maoists, possibly under duress. 36 At the same time, it is also widely believedincluding amongst the Maoiststhat information about specific battles came from Nepal Army officers, e.g. intelligence on Nepal Army positions at Dang was provided to Ananta and Baldev by a Major General who was later elected to the Constitutional Assembly under the Maoist Party. 37 The Maoists themselves claim to have had significant use for 'useful idiots' within the intelligence services of Nepal, particularly the NID. Specifically, for some of the war period the Chief of NID was a former communist party cadre from Rolpa District and under King Gyanendra, he became very close to the Maoists. However, interviews from both sides suggest that the effectiveness of the NID was in question, so the value of the infiltration is difficult to assess. 38 It is, however, very clear that the Maoists did have access to the military and the RNA. Although the actual numbers are difficult to obtain, and the Maoists themselves are unsure of absolute figures, there is clear evidence of excellent access to documentation and approaches such as officer and specialist training. 39 Amongst the infiltration and reliance on local, trusted individual networks, the Maoists also mounted more sophisticated intelligence operations. In Dang I interviewed a former Maoist intelligence Commander, Milan. He had been the chief of intelligence for the district surrounding Dang, a province of significance since it is in the Terai but is also the direct corridor between the Maoist dominated hill regions of Rohlpa and Rukum and the Indian border. 40 Milan was a very sophisticated intelligence operative in Butwal in the Terai. He was a senior Maoist commander who was known (his codename) by the security services. He was a zonal intelligence commander and controlled more than one district and was therefore a leading target for the state security services.
However, during the war he established himself as the head of a small NGO working on migratory links between the hill regions and the Terai. This furnished him with a very plausible alternative identity and also an excuse to travel on his scooter up to Rolpa. His identity also incorporated membership of other political parties, including the NC and UML. He dressed like a western NGO activist and lived a western lifestyle. He lived in a flat close to the local army headquarters where he regularly played volleyball with the officers coming off duty. He also described himself as being 'very friendly' with a number of officers wives. His closest relationship was with an intelligence officer based at HQ who was responsible for tracking down this 'notorious Maoist, Milan'. In fact they discussed various strategies for capturing Milan. 41 Milan was essentially able to embed himself in to the military architecture and provide good quality intelligence using the concentric circles model.

Was it effective?
Overall, the difference in intelligence availability from the local population reflected the overall strategic situation of the war, with the Maoists controlling hill and rural areas and the government (more or less) controlling the Terai and the Kathmandu Valley. This undoubtedly further contributed to the stalemate that eventually developed. However, the Maoists did mount successful intelligence operations in those areas, usually reflecting the concentric circles of activists outlined above. Baldev is a senior Divisional Commander and a close associate of Prachanda (he is one of the core five senior commanders). As such he is one of the very core of Maoists that have been involved from the beginning and are among the most trusted. He was also largely responsible for logistics and intelligence planning throughout much of the war. However, as an illustration of central trust networks in action, he is excellent since he spent two years as a 'businessman' in the Terai running a photographic shop as a cover for an intelligence clearing house. He established a network based on family who came in to the shop and passed on information and was very difficult to break down. Beyond this was a circle of trusted friends and volunteers who provided further information. 42 In many ways, this is illustrative of the trust networks at the heart of Maoist operations. For a start, one would think that someone of the stature of Baldev would be a high risk person to place in a risky situation in the field. However, it does imply that there was a shortage of highly capable intelligence officers that could also be trusted to take independent decisions in the field. The fact that Baldev also had expertise in something saleable like photography also helped with this 'horses for courses' argument.
The Maoists were undoubtedly lucky in facing an inadequate enemy, obsessed with Kathmandu and also lacking understanding of the Nepali countryside. The attitude of the political elite towards the 'backward peasantry' was so entrenched that it was even present in the brahminical approaches to puritan living under the Maoists themselves. 43 Nevertheless the Maoists proved far more able than the government security services in mobilising local communities to act in their own interests, particularly in the base areas where there may have been historical ethnic animosity against the state. The mass recruitment of the Maoists also strengthened some of these relationships. As Nickson 44 pointed out three years before the start of the war itself, the Shining Path had benefitted from a disillusioned aspirational youth who were in or had completed some form of education and did not want to continue as farmers, but had suffered disappointment in pursuing alternative career paths. This class of youth bought in to the alternative vision of a society offered by the Maoists that included opportunities for them to progress in ways that the existing patriarchal and hierarchical system did not, but they also provided a link between the ideology of the Maoists and the families they had come from at the local level. Thus, whilst the leadership may have resembled their opposite numbers, they were able to open channels to the local populations more easily, and to develop relationships whereby they were regarded as less of a threat to civilians than the security services. 45 Again the perceived unpredictable violence of the security services and the fact that they were responsible for the majority of disappearances, undermined government attempts to deal with the local population in many areas and provided additional potential informants for the Maoists. 46 In terms of the organisation and management of Maoist intelligence operations, they relied overwhelmingly on HUMINT, supplemented by use of internet and open source information.
Their two main sources of information were undoubtedly from people on the groundboth their own troops and also civiliansand also through infiltration. Although it is difficult to accurately measure the effectiveness of infiltration, it was undoubtedly successful in terms of gaining an intelligence upper hand over the badly co-ordinated state security institutions. Whilst most of the evidence for this is anecdotal, this is hardly rare within this type of intelligence analysis and it suggests that the Maoists were able to gain traction with individuals up to very high levels of the state institutions themselves. Most of the intelligence was put to the use one might expect in an asymmetrical war, particularly assassination, kidnapping and degradation of state power as well as supporting more conventional warfare like ambushes and attacks on police stations.
The Maoists intelligence structures were centred around a CIB that was a central clearing house for information fed up by the centralised structures to a high command centred around a core of trusted individuals. The Maoist structure was both centralised and decentralised in some ways with some decisions delegated to the local level, particularly to District Commanders, who were able to carry out certain types of operation without permission from the centre. However, these boundaries were very clearly delineated and enforced. The rigours of party discipline were enforced through strict rules and transgressions could be punished severely. Potential transgressions ranged from carrying out actions that did not fit with party policy, or without approval or, usually, where they were held to be for private reasons rather than for the good of the movement. The puritanical streak of the Maoists in upholding what they saw as a form of 'righteousness' could lead to significant punishment for transgressions including demotion, being thrown out of the party and even execution.

International solidarity and the role of India
The international solidarity element is more difficult to get at. Whereas there are commentators who casually claim that the entire peace agreement was 'down to India', usually with no evidence, the actual position of India was far more complex and much more opaque. At various times during the war, India, or parts of the Indian state, local state and some powerful individuals supported almost all groups within the struggle. Officially the Indian Government supported the Nepali state and army, and this was certainly subject to tangible support in terms of arms and munitions and also training.
However, the Maoists were also able to access significant support within India. Several high profile Maoist leaders lived for considerable lengths of time in India openly and were rarely under threat. In interviews with Maoist combatants it also became clear that some had received training within India (although from whom remains somewhat murky) and certainly serious Maoist casualties were treated in Indian hospitals. 47 The relationship between India and Nepal and various Indian organisations, including RAW 48 is complex partly because the relationship between some of the Indian agencies is also not always clear. 49 We will return to this issue below, but in some ways, the Indian position was one of talking to everyone. 50 In other words they placed a bet on everyone in the race rather than just on one horse.
In an interview with a former RAW officer, Jha states that RAW had an ongoing policy of maintaining dialogue with the Maoists even when they were nominally backing the Government in Kathmandu, but this is hardly surprising. 51 However, the palace clearly did not see it that way and spoke of 'doublespeak' on the part of India and even blamed India for covert support for the Maoists. This would certainly chime with the use of the open border for the transport and treatment of Maoist wounded, for example, as well as the lack of action taken against senior Maoist leaders whilst they located in India itself. It should be noted here that several analysts have spent a lot of time trying to ascertain the relationship between the Maoists and India but to little avail. 52 In addition, it is also clear that had the political factions in Nepal been able to work together, then the position of the Indian Government might have been easier. As it was, the different anti-Mao factions spent so much time fighting that the only way to end the conflict was politicallyand that meant making a deal with the Maoists.
In fact, mirroring the disharmony on the part of the anti-Mao movement in Nepal, the Indian Government was doing different things in relation to the Maoists and different elements of Indian Government had different linkages in to Nepali institutions, notably the RAW and the Intelligence Bureau (IB). This relationship came to a head with the arrest of Matrika Yadav, a Madhesi leader at the behest of IB but against the wishes of RAW. That is before we consider any conflict of interest between intelligence officers and diplomats. There is a certain mythology and mysticism about the role of India in Nepal, but in this case it is difficult to ascribe a monolithic policy that defined a constant policy to the Maoists or the Nepali Government over the whole period. As Jha suggests, individuals played a role in how this policy changed over time, suggesting that several RAW officials changed their opinions once they had met with Bhatterai and particularly Prachanda, who clearly impressed them.
However, it is also not entirely true to say that India had no policy on Nepal. India has long history of carrying out similar approaches in Kashmir and Nagaland, which can be characterised as 'engage, promise, coerce, divide, frustrate, corrupt, lure, disappoint', whilst giving up nothing. This was coupled with a policy of supporting Bhatterai as a counter to Prachanda, something that was described to me as a policy of 'always supporting the second strongest person in every movement to keep the leader honest'. 53 What they did in this war was to do this on every side, so they backed every player in a complex and unpredictable game.
Neither is it clear how much support the Nepali Maoists were able to get from similar movements. Shining Path is in Peru and there is no indication that the international revolutionary movements that profess fraternal support provided any practical help, or indeed were in a position to do so. 54 The Naxalites in India, however, were in a better geographical position to help, but again, interviews with former combatants and commanders within Nepal gave rise to no evidence of collusion beyond occasional meetings or liaison officers. Again, given the shortage of resources faced by both the Naxalites and the Maoists it is difficult to determine who was in a position to support whom.
The final point to make about external influences is that the Maoists had very little access to external economic networks that could generate resources. Whereas the Shining Path and Colombian FARC could call on extensive networks of drug trading, the Maoists had to rely on internal sources of income that were largely criminal, including kidnapping, extortion and robbery, including both bank-robbery and theft of private property from 'class enemies'. In addition, once the violence had spread to the Terai, deforestation accelerated rapidly as the Moaists were able to strip out trees and sell the wood over the border in India. Despite this, the Maoists were never able to mobilise sufficient resources to really compete on an equal basis with the Government forces.
Finally, as with anything political in Nepal, one must consider the puzzling position of India during the war. As confusing as a yoga set, the Indian position varied during the war according to which institution was involved at any one time and which individuals were then engaged. Broadly the position shifted from a position at the beginning of supporting a supposedly compliant King, to by the end being at least open to the Maoists. However, in between, India was accused by both sides of being biased and Indian policy cannot be said to have been entirely consistent in taking sides. In the end, the Indian involvement was predominantly arms-length and both supportive and disappointing for both sides. Both sides benefitted from some degree of Indian support at various times, but then also claimed that India supported to other side. The reality was that India was faced with a confusing mass of actors, a complex war and no clear outcome and took the view that firstly, she would support everyone to some degree so that eventually someone sympathetic would win, but, secondly, that she would continue a historical approach of keeping domestic actors weak by supporting rivals within the same party. Thus the Indian intelligence services have tended to support, not the leaders of the political parties, but those who were regarded as the main rivals from the leadership within each party.

Conclusions
Overall, the effectiveness of the intelligence is difficult to measure. One can infer, of course, that it was very good in that the Maoists clearly controlled the countryside and large parts of the urban areas within Nepal. There is also anecdotal information from senior army officers who frequently state that the Maoists knew how many and where any troops were and were ready when they deployed. At the same time, the state knew very little about the Maoist movement itself and was unable to penetrate the Maoist movement itself during the war, suggesting that the overall picture was a mixture of strong Maoist intelligence and weak government intelligence structures.
The Maoist intelligence strategy raises several challenging questions about approaches to intelligence processes. Gentry, proposes an interesting typology for analysing a broad scope of non-state intelligence operations. 55 Intelligence is typically a core party concern in most communist movements and this is clearly reflected in the experience of the Nepali Maoists. Intelligence was integrated to the degree that some senior Maoist cadres integrated intelligence and senior command positions in themselves, although it should be noted that it was uncommon to do both of these things at the same time. Baldev's experience is relevant here since he was the intelligence commander, but within this role he also carried out field operations at particular times as well. In general, the structure of Maoist intelligence surrounded the CIB, which acted as a central training agency, a clearing house for intelligence from the districts and an information processing unit to provide strategic intelligence to the senior command. At the same time, the CIB also enforced strict rules about localised decision-making by District Commanders. 27. I interviewed several officers, both serving and retired for this work, but also in the course of working on demobilisation and transitional justice. However, in their case anonymity is prudent. 28. See note 2 above. 29. Human Intelligence (HUMINT), as opposed to signals intelligence (SIGINT). 30. A former military intelligence source claimed that several Maoists were captured whilst filming during reconnaissance (interview, 2016). 31. Interview with Baldev, Gokarna, 2010. 32. This came up in a discussion with a former Maoist Company Commander in Liwang, 2011. 33. This was verified by interviews both with the Maoists and also with the APF commander in the field at that time. 34. Rosenau 'Subversion and Insurgency'. 35. Rosenau, 'Subversion and Insurgency'. For Northern Ireland there is an extensive literature on intelligence operations by both sides, but for a good insider account of Special Branch activity see Matchett Secret Victory. 36. This information came from an interview with a former Prime Ministerial Adviser to the Nepali Congress during the war. 37. Interviews with senior Maoist commanders, confirmed by military intelligence sources. This person was Kumar Fudong. 38. Opinions on the NID are extremely divided, although interviewees from military intelligence, civilian government and the Maoists were all of the opinion that the NID was ineffective as an intelligence organisation during the war, despite some evidence of UK support (see Bell, Kathmandu). 39. The Maoists had access to significant training materials via this source, as well as early weapons training and advice from former British Ghurkhas. 40. Milan is an interesting character and had used many of his connections to build a very successful property empire. His former comrades all noted that he was now the wealthiest Maoist they knew. At the beginning of the interview, which I partly conducted with Baldev, Baldev himself ordered Milan to tell me the truth and gave him permission to answer my questions. 41. This created much hilarity amongst the Maoists, with Milan being something of a raconteur. The intelligence colonel(who will remain anonymous) went on to become head of military intelligence. 42. I know Baldev well, since he was a leading figure in the peace negotiations. We have been discussing intelligence since 2009. 43. Shah and Pettigrew 'Windows in to a revolution'. 44. Nickson, 'A Peruvian scenario in the making?' 45. See Shah and Pettigrew, 'Windows in to a revolution'. 46. According the Human Rights Commission virtually all human rights crimes reported to them during and since the war have been state perpetrated (interviews, Kathmandu, 2015(interviews, Kathmandu, , 2016(interviews, Kathmandu, , 2017. At the same time, as of time of writing the Truth and Reconciliation Commission has received around 50,000 cases from the war period, around 75% of which are property crimes. Of the remainder, the majority of cases relate to state crimes against humanity rather than Maoist. That is not to say that the Maoists did not perpetrate crimes against humanity, but it is a question of scale. 47. This was also clear when visiting battlefields in the Terai where the Maoists explained how they smuggled the wounded over the border. 48. The Research and Analysis Wing (RAW) or the Indian Secret Service. 49. Jha Battles of the New Republic for an interesting discussion of specific links, including interviews with RAW officers. 50. See Jha Battles of the New Republic. 51. Jha, 'Battles of the New Republic'. This was also supported by field interviews on the Maoist side. 52. It is also worth noting that India was not the only actor with open communication with the Maoistsboth of the NC factions (Deuba and Koirala) kept links open. 53. This was reported to me in an interview with a former NC adviser to the prime Minister. 54. Although Prachanda himself claimed to have a letter of support from Shining Path following correspondence between them. 55. Gentry 'Toward a Theory of Non-State Actors' Intelligence'.

Disclosure statement
No potential conflict of interest was reported by the author.